大型旋转机械非平稳时间序列预测模型的研究  

A study on the prediction model with non-stationary time series to large rotary machinery

作　　者：徐小力[1] 王为真[1] 徐洪安[1] 

机构地区：[1]北京机械工业学院机械工程系,北京100085

出　　处：《北京机械工业学院学报》2001年第3期1-7,共7页Journal of Beijing Institute of Machinery

基　　金：国家自然科学基金项目 [项目编号 :5 9775 0 0 2 ]

摘　　要：目的许多大型旋转机械运行工况恶劣 ,非平稳、非线性特征明显 ,以及各种突发性、偶然性因素的影响 ,给基于振动信号处理的状态预测和状态维护分析带来困难 ,为此研究了非平稳时间序列预测模型。方法在分析传统时序模型振动烈度趋势预测适用性的基础上 ,提出适合于变工况非平稳状态的非平稳时序模型。结果和结论该方法有效地剔除了用于预测的历史数据中的野值及奇异点 。Aim The working conditions of many large rotary machinery are often bad.The characters of non stationary and non linear become more and more remarkable.The effects brought from the unexpected factors and accidental factors in the field of production cannot be ignored, so the analyses of the condition prediction and the condition maintenance based on vibration signal processing become difficult. Therefore a study on the predicting model with non stationary time series is made. Method Based on analyses of the applicability to the traditional time series model for the trend prediction of vibration intensity, a non stationary time series model which adapts to the variable working conditions and non stationary status is put forward. Result and Conclusion The method may pick up the outliers and singular values effectively in the historical data for the prediction and may correct the prediction results.

关 键 词：大型旋转机械 非平衡时间序列 预测模型 

分 类 号：TH17[机械工程—机械制造及自动化] TH113